age and growth of european barbel barbus barbus (cyprinidae) in

Knowledge and Management of Aquatic Ecosystems (2013) 409, 09 http://www.kmae-journal.orgc© ONEMA, 2013

DOI: 10.1051/kmae/2013054

Age and growth of European barbel Barbus barbus(Cyprinidae) in the small, mesotrophic River Leeand relative to other populations in EnglandL. Vilizzi(1)�, G.H. Copp(2),(3), J.R. Britton(3)

Received November 15, 2012

Revised March 18, 2013

Accepted May 16, 2013

ABSTRACT

Key-words:habitat,conservation,scale ageing,recruitment

Suspected of being in decline, the European barbel Barbus barbus popu-lation of the River Lee, a heavily-modified river in South East England, hasbeen the subject of investigations to identify factors associated with per-ceived population decreases. Population surveys between 1995 and 1999captured a total of 912 individuals, and standard length (SL) frequencyanalyses between years suggested that the population decline was notrelated to juvenile recruitment but rather to a recruitment bottleneck in fish300–340 mm SL. This bottleneck probably results from insufficient avail-able habitat suitable to this size class. Of the sampled fish, scales wereremoved from 764 and were used in a scale ageing exercise among threeresearchers. Analyses of their independent age estimates revealed vari-able interpretations, which arose from uncertainties relating to the difficultyof analysing scale patterns from relatively large, slow-growing fish. Never-theless, error was within published acceptable margins, and age estimatesrevealed B. barbus in the river to age 10 years, lower than in many UKrivers. The SL-at-age growth curve was characterised by very fast growthin the initial years of life. Thus, the causal factors in the decline of thisB. barbus population appear to have been in the adult life-stage habitatand were likely related to the loss of longitudinal connectivity, mainly dueto the presence of water retention structures. River and aquatic ecosys-tem remediation strategies should therefore focus on enhancing longitu-dinal connectivity in conjunction with the ongoing improvement of waterquality and ecosystem integrity.

RÉSUMÉ

L’âge et la croissance du barbeau européen Barbus barbus (Cyprinidés) dans la petitemésotrophique rivière Lee et comparaison à d’autres populations en Angleterre

Mots-clés :habitat,conservation,scalimétrie,recrutement

Soupçonnée d’être en déclin, la population de barbeau européen Barbus barbusde la rivière Lee, une rivière fortement modifiée dans le Sud Est de l’Angleterre,a fait l’objet d’investigations pour identifier les facteurs associés à la diminutionressentie de la population. Le suivi de population entre 1995 et 1999 a permis lacapture de 912 poissons, et l’analyse des fréquences de longueur standard (SL)entre années suggère que le déclin de la population n’était pas lié au recrutementdes juvéniles, mais plutôt à un goulot d’étranglement du recrutement chez lespoissons de 300–340 mm SL. Ce goulot d’étranglement résulte probablementd’une insuffisance de l’habitat disponible approprié pour cette classe de taille.

(1) Murray-Darling Freshwater Research Centre, PO Box 991, Wodonga Vic 3689, Australia(2) Salmon & Freshwater Team, Cefas, Pakefield Road, Lowestoft, Suffolk, NR33 0HT, UK, and Environmentaland Life Sciences Graduate Program, Trent University, Peterborough, Canada(3) Centre for Conservation Ecology and Environmental Change, School of Applied Sciences, BournemouthUniversity, Fern Barrow, Poole, Dorset UK� Corresponding author: [email protected]

Article published by EDP Sciences

http://www.kmae-journal.org

http://dx.doi.org/10.1051/kmae/2013054

http://www.edpsciences.org

L. Vilizzi et al.: Knowl. Managt. Aquatic Ecosyst. (2013) 409, 09

Des poissons échantillonnés, les écailles ont été prélevées sur 764 individus etutilisées pour un exercice de détermination de l’âge par trois chercheurs. Les ana-lyses de leurs estimations de l’âge ont révélé des interprétations variables, pro-venant des incertitudes liées à la difficulté d’analyser les écailles de relativementgros poissons à croissance lente. Néanmoins, l’erreur était à l’intérieur des margesacceptables publiées, et les estimations d’âge ont révélé des B. barbus dans la ri-vière jusqu’à 10 ans d’âge, inférieur à celui observé dans de nombreux cours d’eaudu Royaume-Uni. La courbe de croissance SL selon l’âge a été caractérisée parune croissance très rapide au cours des premières années de la vie. Ainsi, lesfacteurs de causalité dans le déclin de cette population de B. barbus paraissentavoir été liés à l’habitat des stades adultes et sont probablement liés à la perte deconnectivité longitudinale, principalement en raison de la présence de structuresde rétention d’eau. Les stratégies de restauration des rivières et des écosystèmesaquatiques devraient donc se concentrer sur l’amélioration de la connectivité lon-gitudinale en collaboration avec l’amélioration permanente de la qualité de l’eauet de l’intégrité de l’écosystème.

INTRODUCTION

The European barbel Barbus barbus (L.) has considerable socio-economic importance as asport fish (Britton and Pegg, 2011) and although is classed on the IUCN Red List as beingof ‘Least Concern’, it is increasingly seen as ‘flag’ species for river conservation. Increasinglythreatened within its native range (Penáz et al., 2003), the IUCN status of B. barbus there-fore carries the caveat of being “locally threatened by water pollution and river regulation”.In the UK, where B. barbus is native to eastern rivers between Yorkshire and the Thames(Wheeler and Jordan, 1990), declining population densities have been reported for some ofthe more urbanised rivers, including the River Lee, Hertfordshire (Pilcher and Copp, 1997;Watkins et al., 1997). River regulation has been identified as a major threat to barbel popula-tions (Pilcher et al., 2004), with compounding adverse effects of elevated nutrient levels on thespecies’ morphology (Tyler and Everett, 1993), physiology (Penáz et al., 2005) and migratorybehaviour (Lucas and Frear, 1997; Vilizzi et al., 2006). This contrasts with earlier, more opti-mistic assessments that suggested that B. barbus numbers were on the increase (Wheelerand Jordan, 1990). Thus, information on population size structure and growth is needed forrivers where B. barbus stocks are thought to be in decline so as to inform environmental man-agers in the formulation of habitat enhancement/rehabilitation and re-stocking strategies thataim to conserve the populations and increase the related angling amenity (Britton and Pegg,2011; Pegg and Britton, 2011).To investigate a perceived decline in the B. barbus population of the River Lee, a smallmesotrophic water course in South East England, collaborative research focused on the pos-sible casual factors (Copp and Bennetts, 1996; Pilcher and Copp, 1997) through studies ofthe species’ environmental biology (e.g. Watkins et al., 1997; Vilizzi and Copp, 2001, 2013a,b;Copp et al., 2002, 2005; Vilizzi et al., 2006). As a component part of this research programme,the aim of the present study was to assess the age and growth rates of River Lee B. barbus.Specific objectives were to: 1) determine inter-annual variation in population size structure; 2)assess seasonal and inter-annual variation in growth; and 3) compare the growth of B. barbusin the River Lee with other stocks throughout England.

MATERIAL AND METHODS

> STUDY AREA

The River Lee (catchment area = 1420 km2; human population ≈ 2 million) is one of the mostheavily human-impacted river systems in the UK, with treated sewage effluent (11 000 m3·d−1)representing up to 80% of the river’s discharge during periods of low flow (NRA 1994a;Faulkner and Copp, 2001), which has consequences for the character of the river substratum(clogging with silt and biofilms, resulting in reduced sediment mobility; Vilizzi et al., 2006).

09p2


A major left-bank tributary of the River Thames, the R. Lee is of chalk stream origin, witha mean annual discharge of 1.09 m3·s−1 (Faulkner and Copp, 2001). The study stretchcomprised three reaches in the upper course of the river, where the river bed slope is≈2.93 m·km−1: 1) from Lemsford Reserve (51.793946N, 0.22895873W) to Stanborough Park(51.787283N; 0.22236586W); 2) from Holwell Bridge (51.772461N; 0.15165210W) to LettyGreen Bridge (51.773503N; 0.14358401W); and 3) from Letty Green Bridge to the easternextent of Woolmers Park (51.771418N; 0.12814522W) (Figure 1).Reach 1 (Lemsford-Stanborough) was about 1800 m in length and was characterised by asubstratum of mainly sand and gravel, with occasional silt and large rocks. From upstreamto downstream, riparian vegetation went from thick and abundant (in the nature reserve) topatchy and relatively thin as the river flows through the park. Channel width varied 6–10 m,with water depths mostly <0.5 m except in one occasional pool (maximum depth = 2 m).Management of the stream course was virtually nil upstream, but manicured downstream inthe park.Reach 2 (Holwell-Letty Green) had a length of ≈1000 m and is situated ≈9 river km down-stream of Reach 1, flows through unfenced, private (sheep) pasture land, where the river sub-stratum varied considerably (i.e. silt and fine sand in back-eddy areas adjacent to pool, withgravel and cobbles in the riffle sections). From upstream to downstream, riparian vegetationwent from a thin strip of trees and bushes to open field with only a few overhanging trees andoccasional reed beds. Water depth varied from shallows of <20 cm to deep (>1.5 m) bends.Management of the stream edges was not apparent, with sheep trampling of the shallowerstream banks.Reach 3 (Letty Green-Woolmers) was about 700 m long and flows through a private estate,beginning upstream with a sinuous section that passes through a small, thick wood, leadinginto a more open section bordered by a few trees, passing under a driveway bridge, whichhas a deep (>2 m) downstream of it. The river then passes into a sequence of runs, pools andriffles that begins within a fenced pasture area (cows) followed by thick woodland in whichthe river is joined by a small, short stream. The riparian border varied from virtually nothing(mostly fencing and tall wild grasses) to >40 m width. Stream width was 3–12 m, with depthsup to 2 m or 3 m. In-stream management had been limited to the removal of fallen trees andbushes (see Copp and Bennetts, 1996).None of the reaches is open to angling activity except at the extreme downstream end ofReach 2. River discharge varied between 0.5 and 5.0 m3·s−1; discharge peaks usually oc-curred in winter and daily flow patterns were strongly influenced by discharges of treated do-mestic effluent (Faulkner and Copp, 2001). However, only the Lemsford Reserve stretch (theclosest to the upstream treatment plant) has occasionally been found to fail non-statutoryriver quality objectives (NRA, 1994b; Copp et al., 2007). Combined with surface run-off fromagricultural land, the water quality of the river is typically mesotrophic. Generally consistentwith the barbel zone (Huet, 1949), other fish species present (in relatively high abundance)were chub Leuciscus cephalus, European minnow Phoxinus phoxinus, gudgeon Gobio go-bio, stone loach Barbatula barbatula, European bullhead Cottus gobio and dace Leuciscusleuciscus, with occasional occurrences of roach Rutilus rutilus, perch Perca fluviatilis, three-spine stickleback Gasterosteus aculeatus and (stocked) brown trout Salmo trutta (Copp andBennetts, 1996; Pilcher and Copp, 1997; Copp et al., 2005). In this part of the river, ben-thic invertebrates have been found to be dominated by Baetis rhodani, Ephemerella ignitaand Chironomidae larvae (Copp et al., 2005), and drifting invertebrates by Caenis robusta, B.rhodani, E. ignita, Asellus aquaticus and Chironomidae (Edmonds-Brown et al., 2004).

> FISH COLLECTION AND MEASUREMENT

Barbus barbus were collected during 58 scheduled and incidental sampling excursions (Vilizziet al., 2006) between August 1995 and April 1999. Fish were captured using a back-pack elec-trofishing unit (Deka 3000: Deka-Geratebau, Rudolf Muhlenbein, VincentiusstraBe 13, 3538Marsberg, Germany) whilst fishing in an upstream direction. Previous depletion sampling

09p3


revealed the maximum standard length (SL) size of B. barbus in Reach 3 to be of 450 mm(Copp and Bennetts, 1996), and more widely in the River Lee rarely exceed 500 mm (G.H.Copp, unpublished data). Upon capture, fish were measured (nearest mm) for SL, fork (FL)and total length (TL), weighed (nearest g) and examined for sex (using external features andrunning sexual products during the reproductive season only, typically May to July). Up to fiveun-regenerated scales were removed from the antero-medial region of the body immediatelybelow the lateral line (Penáz and Štouracová, 1991), cleaned of mucus and stored in pa-per envelopes for subsequent examination. In the laboratory, scales were mounted betweenglass slides after soaking briefly in 20% sodium hypochlorite and rinsing in distilled water.The criteria of Chugunova (1963) were followed for annulus identification, with the scalesinspected with a dissecting microscope (10−20×) under transmitted light. An annulus wastaken as a transition between two uninterrupted zones of closely- and widely-spaced circuli,with anastomosis as an essential criterion for identification. To ensure the consistency of ageinterpretation, three interpreters (A, B and C) counted annuli on the scales under study. When-ever the pattern of closely- and widely-spaced circuli could not be consistently identified thepreparation was scored as ‘uninterpretable’. The interpreters were given no indication of thedate of capture, length or weight of the fish.

> DATA ANALYSIS

Inter-annual variation in size structure was examined by plotting of length-frequency distribu-tions from 1995 to 1998, with each year taken to represent the ‘growth season’ of the three‘biological seasons’ (i.e. reproduction, growth, over-winter) as described by Cattanéo (2005).For B. barbus, the growth season (i.e. when fish are feeding actively) was nominally taken asstarting on June 1. Annual size structures were summarised using proportional stock density(PSD), which was calculated as the ratio of quality-sized fish to stock-sized fish (cumulativesums) multiplied by 100. The following size classes (TL) were defined (Abernethy, 2006): (i)“Zero” fish (<280 mm) have little recreational value being smaller than stock size, but eco-logically represent an indicator of successful recruitment (i.e. young-of-year and immatures,noting that male barbel generally mature at a smaller size than females); (ii) ‘Stock’ fish (280–410 mm) have a minimum size of 20–26% of the world-record length; (iii) ‘Quality’ fish (410–530 mm) fall into the 36–41% range and are generally of appeal to anglers; (iv) ‘Preferred’(530–660 mm; 45–55%) are of even higher appeal; (v) ‘Memorable’ (660–840 mm; 59–64%)is the size most anglers remember catching; and (vi) ‘Trophy’ (>840 mm; 74–80%) is a sizeconsidered worthy of acknowledgement. Notably, the percentage ranges for each size classcorrespond to appropriately defined portions of the Weithman’s sigmoid curve (Gabelhouse,1984). For each year of sampling, PSD was computed in the R language and software envi-ronment v2.14.2 (R Development Core Team, 2010) using libraries FSA and NCStats (Ogle,2011a). Given the similarity in maximum reported size for B. barbus and common carp Cypri-nus carpio carpio (www.fishbase.org), the size classes for the latter species as provided inOgle (2011a) were extrapolated to B. barbus.Bias and precision of annulus counts were analysed by statistical and graphical methodsfor each pair of replicate counts from the three interpreters. Accordingly, an age bias plot(Campana et al., 1995) was generated and a test of symmetry (Hoenig et al., 1995) computedfrom the corresponding age-agreement table. In this table, the main diagonal represents thefrequency of fish for which the same estimated age is obtained by any combination of twointerpreters, whereas each cell off the main diagonal represents a difference in estimated agebetween the interpreters. If there are no systematic differences between interpreters, then itis expected that the disagreements in estimated age would fall randomly on either side of thediagonal, producing an approximately symmetric age-agreement table. Precision, defined asthe reproducibility of estimated ages between interpreters (Campana, 2001), was computedas percentage agreement, average percent error (APE), and by the coefficient of variation (CV).All computations were in R using libraries FSA and NCStats and following guidelines in Ogle(2011b).

09p4

www.fishbase.org


Growth of B. barbus in the River Lee was described in terms of weight-SL (WL), SL-at-age(LAA), and weight-at-age (WAA) relationships both for the entire stock and for males andfemales separately. For computation of LAA and WAA relationships, the mean estimated agefrom the three interpreters was adjusted by taking June 1 as the nominal hatching date ofeach annual cohort in the sampled stock, with the final estimated age of each fish expressedas a decimal:

age (as integer) + (days from June 1 to date of capture) / 365.25.

Fitting of WL relationships and of the von Bertalanffy growth model (VBGM) to LAA and WAAdata was in R using libraries FSA, NCStats and nlstools and following guidelines in Ogle(2011c,d). For the VBGM, which has been found to provide unrealistic estimates of ultimatesize (Živkov et al., 1999; Copp et al., 2004), both the Beverton-Holt and Gallucci-Quinn pa-rameterisations were used, the former providing estimates of SLinf (or Winf), k and t0 for com-parisons with other studies and the latter facilitating between-sex comparisons by means ofparameter ω = FLinf · k (or Winf · k), which has been shown to capture the essential featuresof growth and to facilitate growth-rate comparisons (Charnov, 2010). For WAA growth com-putations, the Beverton-Holt adaptation of the VBGM was used (Ricker, 1975), but with theslope parameter fixed at b = 3 due to computational constraints. Following Ogle (2011d),between-sex growth comparisons were by hierarchical fitting of eight VBGMs in total underboth parameterisations as follows: (i) a general model with separate parameter estimates forindividuals of each sex; (ii) three models with one parameter in common between sexes; (iii)three models with two parameters in common between sexes; and (iv) one common modelwith the same parameter estimates for both sexes (also to account for uncertainty in sex-ing some individuals). After fitting of the individual models, the Akaike (AIC) and BayesianInformation Criterion (BIC) were computed to select the best-fitting model, with preferencegiven to BIC in case of disparity of outcomes for reasons of model parsimony (Burnham andAnderson, 2003). Notably, SL was used as the reference measure for length in all growthmodels. This was to facilitate comparisons with other B. barbus stocks (see below) and be-cause SL (or, equivalently, FL) measurements are not affected by possible damages or erosionof the lobes of the caudal fin.Mean LAA values (combined sexes) for B. barbus stocks from rivers throughout England werecollated from literature as well as unpublished data, with FL converted into SL whenever ap-plicable (conversion factor as per www.fishbase.org). Both the VBGM and a log-log quadraticmodel (LLQM) were fitted to the LAA data for each stock (cf. Vilizzi and Walker, 1999; Brittonet al., 2012). Fitting of the VBGM was in FiSAT II v1.2.2 (www.fao.org) using a maximumSL value of 1000 mm (maximum reported length for B. barbus; anonymous reviewer, pers.comm.) and with a starting t0 value ranging from −2 to 2 (or fixed at 0 in case of lack of fit).Whenever the VBGM could not be fitted because of less than seven LAA values in a stock,or due to linear (i.e. non-asymptotic) growth, the stock was excluded from analysis. Also, forconsistency with the other stocks the VMGM was re-fitted to the River Lee B. barbus stock ofthe present study based on the mean LAA values. Fitting of the LLQM was in S-Plus 6 Profes-sional for Windows R©. The resulting VBGM and LLQM parameters for the different B. barbusstocks were then subjected separately to principal component analysis (PCA) following nor-malisation (PRIMER v6: Clarke and Gorley, 2006). Finally, for the VBGM the growth param-eters ϕ′ and ω were also computed for each stock (Živkov et al., 1999; Charnov, 2010), andan overall VBGM curve was fitted to the LAA data from all available B. barbus stocks fromEngland (i.e. including those for which a VBGM and/or a LLQM could not be fitted because ofthe above constraints) using the same fitting procedures as above for the R. Lee population.

RESULTS

In total, 912 B. barbus were captured between 1995 and 1998, of which 148 from Reach 1,396 from Reach 2, and 368 from Reach 3. The largest size achieved was 500 mm SL, thelargest weight 2000 g, and the oldest estimated age 10 years (sex unknown).

09p5

www.fishbase.org

www.fao.org


Figure 1The three study reaches on the River Lee (Hertfordshire, England): Reach 1, extending downstream fromLemsford Reserve to Stanborough Park; Reach 2, extending from Holwell Bridge to Letty Green Bridge;and Reach 3, extending from Letty Green Bridge to the eastern extent of Woolmers Park.

Table IInter-annual variation in size structure (total number and relative percentage of fish) and proportionalstock density (PSD, with 95% confidence intervals) of European barbel Barbus barbus in the River Lee.Z = Zero; S = Stock; Q = Quality; P = Preferred; M = Memorable; T = Trophy (see text for definition ofsize classes). See also Figure 2.

Size classYear n Z S Q P M T PSD1995 196 76 (39) 35 (18) 81 (41) 4 (2) 0 (0) 0 (0) 71 (62–78)1996 70 46 (64) 6 (8) 18 (25) 2 (3) 0 (0) 0 (0) 75 (55–88)1997 321 107 (33) 79 (24) 135 (41) 8 (2) 0 (0) 0 (0) 63 (56–69)1998 304 110 (36) 66 (21) 128 (41) 5 (2) 0 (0) 0 (0) 66 (59–72)

Length-frequency distributions for the four years of study were distinctively bimodal due toa predominance of fish in the smaller (Zero) and larger (Quality) size classes and a gap inthe middle-sized (Stock) class (Figure 2). This pattern was reflected by the relative proportionof size classes in the R. Lee population (Table I), with Quality fish being more abundant andpresent in identical proportions in all years except 1996, when Zero fish were predominant.Preferred fish always represented a minor proportion of the catch, and neither Memorable norTrophy fish were ever caught. Finally, PSD was similar in 1995 and 1996, and slightly higherin those years relative to 1997 and 1998 (Table I).

09p6


Figure 2Size structure of B. barbus in the River Lee over four years of study with indication of size classes andtheir annual progression (0+, 1+, 2+, 3+ = age classes). Z = Zero; S = Stock; Q = Quality; P = Preferred.See also Table I.

09p7


Table IIBias (test of symmetry) and precision of age estimates for B. barbus in the River Lee based on scaleannulus counts from three interpreters. APE = Average Percentage Error; CV = Coefficient of Variation.Significant P values in bold (α = 0.05). See also Figure 3.

Bias PrecisionInterpreters n df χ2 P Agreement APE CVA vs. B 673 28 36.87 0.122 46.51 8.49 12.01A vs. C 687 27 52.22 0.002 44.25 9.11 12.88B vs. C 687 26 46.05 0.009 55.36 6.57 9.29

In total, 764 scales were examined for annulus counts. The proportion of scales discarded asuninterpretable was 8.2% (n = 58), 12.4% (n = 84) and 10.1% (n = 70) for interpreters A,B and C, respectively. Pair-wise bias comparisons showed a tendency for interpreters Band C to over-estimate annulus counts beyond seven annuli (Figure 3), even though thesedifferences were statistically significant only relative to interpreter C, hence not between in-terpreters A and B (test of symmetry: Table II). Conversely, agreement was higher betweeninterpreters B and C, whose corresponding APE and CV values were lower relative to inter-preter A vs. B and C (Table II). For this reason, choice of assigning a mean estimated age tofish based on annulus counts from all three interpreters (rather than age estimates based onone interpreter only) was deemed appropriate.The mean estimated value for the coefficient b in the WL relationship was <3, and upon com-parison between sexes there was no significant difference in slopes (P > 0.05), although theintercept for males was larger (P < 0.001) than for females (Table III, Figures 4a, 4b). Growthin SL for the entire stock was asymptotic and differed between males and females, with thelatter attaining a larger asymptotic SL, which was reflected by a higher growth parameter ω(Table III, Figures 4c, 4d). Growth in weight for the entire stock also was asymptotic anddiffered between males and females, the latter attaining a higher asymptotic weight, whichwas faster in early years of life (as indicated by parameter k) compared to males (Table III;Figures 4e, 4f).For inter-population comparisons, mean LAA data were available for 25 English B. barbusstocks in total, with a maximum estimated age (scale annulus counts) of 15 years (Table IV).A VBGM could be fitted to 19 of the stocks under study, whereas a LLQM to 23 of them(Table V). Compared to other stocks, growth of B. barbus in the R. Lee was very fast inthe first two years of life compared to most of the other stocks and was similar to that inthe Severn (1974 and 2012) (Figure 5), even though growth parameters ϕ′ and ω fell withinthe range of the stocks under study (Table V). The largest SLinf value was achieved by theBristol Avon stock, and the largest k (and ω) by the 2012 Severn stock (Table V; Figure 5).The overall VBGM for the 25 English B. barbus stocks resulted in the growth parametersSLinf = 567.4 (±21.3 SE), k = 0.156 (±0.014 SE) and t0 = 0.518 (±0.140 SE) (Figure 6).

DISCUSSION

The size-class distributions of B. barbus in the R. Lee during the four years of study revealeda chronological progression of young fish through the Zero class in 1995 to 1997, and be-ginning to enter the Stock class by 1998 (Fig. 2). The Quality class of larger fish remainedrelatively constant, albeit with a dip in numbers overall in 1996. The apparent gap around300–340 mm TL at the lower end of the Stock class resembles a recruitment bottleneck sim-ilar to that described for nase Chondrostoma nasus in regulated stretches of the upper RiverRhône (Persat and Chessel, 1989). Similar to B. barbus, C. nasus is a large-bodied benthicspecies, and a gap in recruitment between smaller and larger size classes (extending overa range of 600 mm) was attributed to insufficient suitable habitat in the by-passed, naturalstretches of the river, where a minimum discharge regime is in place for hydroelectric pro-duction (Persat and Chessel, 1989). In the upper R. Lee, the main channel is not by-passed,

09p8


Figure 3Age-bias plots with pair-wise comparisons of B. barbus scale annulus counts from three interpreters.See also Table II.

09p9


Tab

leIII

Par

amet

eres

timat

es(w

ith95

%lo

wer

and

upp

erco

nfid

ence

inte

rval

s:LC

Iand

UC

I,re

spec

tivel

y)fo

rw

eigh

t-le

ngth

(WL)

,len

gth-

at-a

ge(L

AA

)(st

and

ard

leng

th:S

L,m

m),

and

WA

Are

latio

nshi

ps

forB

.bar

bus

inth

eR

iver

Lee.

For

the

von

Ber

tala

nffy

grow

thm

odel

(VB

GM

),th

eB

ever

ton-

Hol

t(B

-H)a

ndG

allu

cci-

Qui

nn(G

-H)p

aram

eter

isat

ions

are

give

nfo

rm

ales

and

fem

ales

.See

also

Figu

re4.

WL

VB

GM

(LA

A)

VB

GM

(WA

A)

Par

amet

erE

stim

ate

SE

LCI

UC

IP

aram

eter

Est

imat

eS

ELC

IU

CI

Par

amet

erE

stim

ate

SE

LCI

UC

IO

vera

llO

vera

llO

vera

lla

−9.8

10.

07−9

.95−9

.68

SL i

nf59

1.0

36.7

528.

369

5.3

Win

f12

3669

1131

1430

b2.

790.

012.

762.

81k

0.13

90.

016

0.10

60.

172

k0.

364

0.03

00.

286

0.41

0t 0

−0.0

810.

129−0

.360

−0.1

33t 0

1.23

60.

163

0.82

51.

445

Sex

Sex

Sex

a M−9

.17

0.38

−9.9

3−8

.41

B-H

B-H

a F−9

.09

0.38

−9.9

0−8

.28

SL i

nfM

404.

320

.737

5.5

459.

8W

inf M

815

7569

6.3

1025

.2b

2.67

0.06

2.56

2.66

k0.

325

0.07

00.

195

0.45

8k M

0.61

20.

152

0.35

70.

965

t 00.

892

0.42

7−0

.221

1.54

3t 0

1.82

90.

460

0.77

82.

597

SL i

nfF

441.

522

.040

9.4

502.

1W

inf F

1291

138

1076

1655

k F0.

418

0.08

60.

289

0.63

2G

-QG

-Qω

M13

1.3

22.0

86.9

177.

7ω

574.

655

.748

5.8

702.

6k

0.32

50.

070

0.18

30.

470

k M0.

739

0.09

30.

565

0.95

9t 0

0.89

20.

427−0

.334

1.51

0t 0

2.10

30.

320

1.38

62.

694

ωF

143.

424

.095

.619

3.8

k F0.

461

0.07

70.

325

0.63

7

09p10


Figure 4Growth relationships for B. barbus in the River Lee with estimated ages from scale annulus counts. (a,b) Weight-standard length relationship for the entire stock and for males and females separately; (c, d)Standard length-at-age based on the von Bertalanffy growth model (VBGM) for the entire stock and formales and females separately; (e, f) Weight-at-age based on the VBGM for the entire stock and for malesand females separately. See also Table III.

09p11


Tab

leIV

Mea

nLA

Ava

lues

(SL,

mm

)for

Eng

lish

B.b

arb

usst

ocks

.Est

imat

edag

esar

eb

ased

onsc

ale

annu

lus

coun

ts.R

efer

ence

s(R

ef.)

are:

1)H

unt

and

Jone

s(1

975)

,2)T

aylo

ret

al.(

2004

),3)

Brit

ton

etal

.(20

12),

4)un

pub

lishe

d,a

nd5)

pre

sent

stud

y.

Est

imat

edag

eID

Riv

er1

23

45

67

89

1011

1213

1415

Ref

.1

Sev

ern

(197

1)50

115

166

216

270

298

340

399

433

502

528

12

Sev

ern

(197

2)55

8612

515

819

121

724

225

927

827

628

930

91

3A

rrow

4091

133

216

291

330

328

380

370

398

398

408

502

24

Der

wen

t50

8516

722

220

335

140

748

248

849

150

541

241

944

42

5N

idd

3676

315

329

357

430

455

469

418

419

26

Tem

e83

209

330

371

349

424

411

488

461

449

479

27

Tham

es42

116

194

216

415

466

471

493

532

564

28

Bris

tolA

von

4193

145

186

206

309

367

435

459

470

39

Dan

e44

9914

018

625

429

829

933

03

10D

on14

621

425

128

335

941

440

750

443

946

150

349

750

33

11Iv

el48

9013

422

329

642

939

746

542

950

052

756

955

461

23

12K

enne

t90

153

182

236

310

405

366

387

428

461

480

525

526

516

313

Lod

don

9815

231

342

741

945

844

347

044

244

155

23

14M

edw

ay37

110

161

207

361

422

476

555

515

315

Mol

e46

8115

120

727

934

739

443

143

338

83

16R

othe

r48

183

188

289

327

341

370

391

421

399

407

317

Soa

r50

8918

824

729

937

143

450

245

449

945

452

956

93

18W

ey53

9316

726

529

737

141

043

445

044

648

945

953

052

543

23

19W

harf

e40

198

256

283

341

368

428

320

Aln

e20

622

755

540

84

21C

ald

er21

825

528

229

833

937

04

22D

earn

e20

936

437

240

242

04

23S

ever

n(2

012)

155

356

454

490

432

442

432

424

WA

von

306

546

428

425

Lee

112

178

206

285

331

363

381

400

396

455

5

09p12


Tab

leV

VB

GM

(SL,

mm

)and

log-

log

qua

dra

ticm

odel

(LLQ

M)p

aram

eter

s(w

ithst

and

ard

erro

rs:S

E)f

or23

Eng

lish

B.b

arb

usst

ocks

(IDfr

omTa

ble

IV).

See

also

Figu

re5.

VB

GM

LLQ

MID

Riv

erS

L in

fS

Ek

SE

t 0S

Eϕ′

ωa

SE

bS

Ec

SE

1S

ever

n(1

971)

––

––

––

––

0.43

30.

118

1.44

60.

156−0

.161

0.05

02

Sev

ern

(197

2)–

––

––

––

–0.

694

0.15

01.

184

0.19

5−0

.151

0.06

23

Arr

ow51

756

0.17

0.04

0.70

0.35

4.66

87.9

−0.4

340.

276

2.51

70.

355−0

.497

0.11

14

Der

wen

t50

652

0.22

0.07

0.93

0.48

4.75

111.

2−0

.274

0.45

42.

388

0.56

9−0

.468

0.17

25

Nid

d49

010

20.

240.

100.

900.

344.

7611

7.6

−0.9

390.

472

3.12

40.

636−0

.666

0.20

46

Tem

e51

443

0.18

0.05

1.04

0.48

4.68

92.5

−1.4

410.

403

3.63

00.

467−0

.799

0.13

37

Tham

es71

780

0.13

0.03

0.63

0.24

4.82

93.2

−0.4

030.

199

2.52

60.

261−0

.486

0.08

28

Bris

tolA

von

836

268

0.07

0.04

0.41

0.40

4.69

58.5

0.20

00.

213

1.67

40.

277−0

.213

0.08

79

Dan

e49

414

00.

150.

070.

500.

344.

5674

.10.

159

0.26

21.

759

0.36

9−0

.268

0.12

610

Don

557

630.

190.

051.

490.

544.

7710

5.8

−1.5

920.

613

3.73

40.

670−0

.808

0.18

111

Ivel

748

950.

110.

030.

660.

344.

7982

.30.

041

0.28

31.

897

0.35

7−0

.291

0.10

812

Ken

net

640

690.

130.

030.

930.

434.

7383

.2−0

.696

0.40

52.

717

0.46

4−0

.525

0.13

113

Lod

don

540

670.

180.

071.

070.

594.

7297

.2−0

.869

0.55

63.

013

0.65

4−0

.631

0.18

714

Med

way

728

185

0.12

0.05

0.71

0.41

4.80

87.3

−0.3

920.

237

2.42

10.

311−0

.442

0.09

815

Mol

e54

314

50.

190.

100.

850.

494.

7510

3.1

0.05

70.

466

1.83

70.

629−0

.263

0.20

616

Rot

her

475

450.

200.

050.

550.

334.

6595

.0−0

.239

0.24

72.

428

0.32

4−0

.503

0.10

317

Soa

r60

051

0.17

0.03

0.73

0.30

4.79

102.

0−0

.301

0.29

72.

436

0.37

5−0

.474

0.11

518

Wey

529

310.

220.

050.

820.

304.

7911

6.3

−0.4

000.

303

2.61

40.

379−0

.543

0.11

519

Wha

rfe

547

570.

180.

040.

550.

174.

7398

.4−0

.844

0.16

03.

180

0.22

5−0

.720

0.07

721

Cal

der

––

––

––

––

2.40

80.

373−0

.369

0.47

90.

246

0.15

222

Dea

rne

––

––

––

––

−0.4

050.

588

3.19

60.

764−0

.844

0.24

423

Sev

ern

(201

2)45

854

0.35

0.21

−0.1

60.

834.

8716

0.5

0.99

50.

284

1.56

30.

394−0

.368

0.12

724

Lee

518

640.

180.

05−0

.27

0.40

4.68

93.2

1.19

90.

212

0.96

50.

286−0

.119

0.09

4

09p13


Figure 5Principal component analysis (PCA) plot for B. barbus stocks in England based on the VBGM (top) andLLQM (bottom) parameters (cf. Table V).

but it is interrupted throughout its course by water retention structures (i.e. weirs and reser-voir dams), which result in B. barbus movements during normal discharge conditions beingrestricted to the confines of consecutive weirs (Vilizzi et al., 2006). Tagged B. barbus in aconcurrent study (Vilizzi et al., 2006), which encompassed Reaches 2 and 3, have been re-captured both upstream and downstream of the weirs that delimit this stretch of river; driftingB. barbus eggs and larvae have also been reported for this same stretch of river (Copp et al.,2002). These studies suggest that likely exceptions to confined movement include the down-stream displacement of small B. barbus (over the water retention structures) during virtuallyall discharge rates, and the upstream movement by some larger B. barbus during extremehigh discharge events (evinced by angler captures of tagged barbel; Vilizzi et al., 2006).As such, the present study demonstrates that recruitment problems in the R. Lee may not bedue to a shortage of juvenile habitat, as originally perceived (Pilcher and Copp, 1997; Watkinset al., 1997; Copp et al., 2002), but rather to some other factors that affect the intermediate-sized fish, the most likely being the impediment to free movement imposed by water retentionstructures. Similar to C. nasus in the Upper Rhône (Persat and Chessel, 1989), the recruitmentgap in B. barbus of the R. Lee could be due to larger fish occupying the entire ‘suitable’ habi-tat, leaving none for up-coming recruits, which are forced to move elsewhere in the system.

09p14


Figure 6Overall VBGM for 25 stocks of B. barbus in England (cf. Table IV).

Indeed, the sensitivity of larger barbel (i.e. >100 mm SL) to reduced habitat was demon-strated in Reach 3, where in an earlier study a significant decrease in mean SL (from 330to 202 mm) was observed following removal of in-stream and over-hanging branches fromthat reach (Copp and Bennetts, 1996). Alternatively, this gap could be an artefact of inter-annual variations in recruitment or that of the discontinuous nature of this river system, whichinhibits free movement of B. barbus, and thus biasing perception of the overall populationstructure.The present study has evaluated bias and precision of age estimates in B. barbus, eventhough Britton et al. (2012) utilised a quality control procedure to minimise errors in age es-timation based on the evaluation of a set of scales from each individual fish. Evaluation ofbias and precision in fish age-growth studies is a valuable means, inter alia, of comparingthe skill level of one interpreter relative to that of others (Campana, 2001), and the presentstudy indicated a tendency for two (i.e. B and C) of the three interpreters to count more annulirelative to those counted by the other interpreter (A). A reason for this discrepancy could beattributable to the greater experience in ageing fish by Interpreter A (LV) relative to B and C.On the other hand, it can be argued that for the same reason Interpreter A may have beeninclined to take a more conservative approach due to concerns with the limitations of scalesrelative to other structures in ageing older fish (e.g. Vilizzi et al., 1998; Vilizzi and Walker,1999). This assumption for B. barbus remains conditional on future comparative studies thatare based on different ageing structures. Clearly, validation studies of the ages of B. barbus inthe R. Lee will help resolve these interpretability issues, even though always within the con-straints of age estimates based on non-destructive techniques (e.g. removal of scales), asopposed to methods involving the sacrifice of fish (e.g. otolith extraction), due to the conser-vation value of B. barbus populations in England (Britton and Pegg, 2011; Pegg and Britton,

09p15


2011). Finally, the APE and CV values achieved in the present study, albeit higher than thosereported as ‘desirable’ in the age-growth literature (i.e. APE < 5.5% and CV < 7.6%), werestill in accordance with those reported for scales (Campana, 2001).The importance of precision in the scale age estimates relates to their role in assessing theage and growth rates of the R. Lee B. barbus in relation to other populations. Individualswere aged to age 10 and although this suggests that B. barbus were capable of a relativelylong life span in the river, individual barbel in other English rivers have been recorded to beat least 20 years old (Britton et al., 2012). Given the number of individual fish sampled inthe R. Lee, it was assumed that no age classes were missed and so 10 years is close totheir maximum attainable age under those river conditions. The SL-at-age data and growthmodels from the R. Lee were in accordance with other UK riverine populations and alsothose in Europe (e.g. Przybylski et al., 2004; Prokeš et al., 2006). Nevertheless, the SL-at-agedata suggested a large growth increment in the first year of life compared with other rivers,with slowing thereafter and producing a maximum theoretical length (518 mm) that whilst inthe range of other populations, may still be considered relatively low. Similar to the factorsaffecting B. barbus recruitment outlined above, these growth characteristics were acquiredfrom three reaches within a much larger river basin but may be considered symptomatic ofthe environmental conditions of the habitats and their disconnection, i.e. favourable in theinitial years of life for B. barbus but potentially limiting thereafter.In summary, the present study reveals that the declining B. barbus stocks in the R. Lee wasprobably related to a recruitment bottleneck surrounding the progression of the 300–340 mmSL class into larger size classes, most probably linked to insufficient available habitat for fishof that size range and in the absence of known stocking (M. Carter, pers. comm.). Juvenilegrowth and recruitment does not appear to be problematic and thus options for improvedmanagement of the aquatic ecosystem from a B. barbus perspective should consider aspectsof improved longitudinal connectivity and adult habitat enhancement, rather than focusing onspawning and nursery areas.

ACKNOWLEDGEMENTS

We thank the Hertfordshire & Middlesex Wildlife Trust, Welwyn Garden City Council, the CecilGascogne Estates, and the late Mrs Lucas for permission to carry out our investigations ontheir stretches of the River Lee. We also thank, T. Bennetts, J. Cerný, R. Cortes, S. Doherty,K. Haynes, V. Jubb, J. Majecki, S. Newby, W. Pegg, M. Watkins, K.J. Wesley, and the lateS. Smith for assistance in the field.

REFERENCES

Abernethy D.L., 2006. An angler’s guide to interpreting Alabama wildlife and freshwater fisheries reser-voir reports. Department of Conservation and Natural Resources, Wildlife and Freshwater FisheriesDivision, Fisheries Section, Alabama USA. http://www.outdooralabama.com/fishing/freshwater/where/reservoirs/guide.pdf

Britton J.R. and Pegg J., 2011. Ecology of European barbel Barbus barbus: implications for river, fishery,and conservation management. Rev. Fish. Sci., 19, 321–330.

Britton J.R., Davies G.D. and Pegg J., 2012. Spatial variation in the somatic growth rates of Europeanbarbel Barbus barbus: a UK perspective. Ecol. Freshwat. Fish, 22, 21–29.

Burnham K.P. and Anderson D.R., 2003. Model selection and multimodel inference. a practicalinformation-theoretic approach. New York, Springer.

Campana S.E., 2001. Accuracy, precision and quality control in age determination, including a review ofthe use and abuse of age validation methods. J. Fish Biol., 9, 197–242.

Campana S.E., Annand M.C. and McMillan J.I., 1995. Graphical and statistical methods for determiningthe consistency of age determinations. Trans. Am. Fish. Soc., 124, 131–138.

09p16

http://www.outdooralabama.com/fishing/freshwater/where/reservoirs/guide.pdf

http://www.outdooralabama.com/fishing/freshwater/where/reservoirs/guide.pdf


Cattanéo F., 2005. Does hydrology constrain the structure of fish assemblages in French streams? Localscale analysis. Arch. Hydrobiol., 164, 345–365.

Charman D.J., 2010. Centennial climate variability in the British Isles during the mid-late Holocene.Quater. Sci. Rev., 29, 1539–1554.

Charnov E.L., 2010. Comparing body-size growth curves: the Gallucci-Quinn index, and beyond.Environ. Biol. Fish., 88, 293–294.

Chugunova N.I., 1963. Age and growth studies in fish. Office of Technological Services, Washington,D.C.

Clarke K.R. and Gorley RN., 2006. PRIMER v6: User Manual/Tutorial. PRIMER-E Ltd, Plymouth, UK.

Copp G.H. and Bennetts T.A., 1996. Short-term effects of removing riparian and instream cover onbarbel (Barbus barbus) and other fish populations in a stretch of English chalk stream. Folia Zool.,45, 283–288.

Copp G.H., Doherty S., Faulkner H., Watkins M.S. and Majecki J., 2002. Diel drift behaviour of fish eggsand larvae, in particular barbel, Barbus barbus (L.), in an English chalk stream. Fish. Manag. Ecol.,9, 95–104.

Copp G.H., Fox M.G., Przybylski M., Godinho F. and Vila-Gispert A., 2004. Life-time growth patterns ofpumpkinseed Lepomis gibbosus introduced to Europe relative to native North American popula-tions. Folia Zool., 53, 237–254.

Copp G.H., Spathari S. and Turmel M., 2005. Consistency of diel behaviour and interactions of streamfishes and invertebrates during summer. Riv. Res. Appl., 21, 75–90.

Copp G.H., England J., Tyner R., Carter M.G., Przybylski M. and Wesley K.J., 2007. Growth and con-dition of dace Leuciscus leuciscus in the River Lee (Hertfordshire) relative to selected populationselsewhere in Europe. London Nat., 86, 71–85.

Edmonds-Brown V., Copp G.H. and Majecki J., 2004. Diel patterns of drift by macroinvertebrates in theRiver Lee (Hertfordshire) during low discharge. London Nat., 83, 145–157.

Faulkner H. and Copp G.H., 2001. A model for accurate drift estimation in streams. Freshwater Biol.,46, 723–733.

Gabelhouse D.W., 1984. A length-categorization system to assess fish stocks. North Am. J. Fish.Manage. 4, 273–285.

Hoenig J.M., Morgan M.J. and Brown C.A., 1995. Analysing differences between two age determinationmethods by tests of symmetry. Can. J. Fish. Aquat. Sci., 52, 364–368.

Huet M., 1949. Aperçu de la relation entre la pente et les populations piscicoles des eaux courantes.Schweiz. Z. Hydrol., 11, 332–351.

Hunt P.C. and Jones J.W., 1975. A population study of Barbus barbus L. in the River Severn, England.III. Growth. J. Fish. Biol., 7, 361–376.

Lucas M.C. and Frear P.A., 1997. Effects of a flow-gauging weir on the migratory behaviour of adultbarbel, a riverine cyprinid. J. Fish Biol., 50, 382–396.

NRA, 1994a. Upper Lee catchment management plan: Final plan. National Rivers Authority, ThamesRegion, Reading, UK

NRA, 1994b. Implementation of the Freshwater Fish Directive: water quality requirements for the supportof fish life. National Rivers Authority, Rivers House, Waterside Drive, Aztec West, Almondsbury,Bristol, England.

Ogle D., 2011a. FishR vignette – Size structure analysis. Northland College, Wisconsin, USA, 16 p.,http://www.ncfaculty.net/dogle/fishR/gnrlex/SizeStructure/SizeStructure.pdf.

Ogle D., 2011b. FishR vignette – Precision and accuracy in ages. Northland College, Wisconsin, USA,11 p., http://www.ncfaculty.net/dogle/fishR/gnrlex/AgeComparisons/AgeComparisons.pdf.

Ogle D., 2011c. FishR vignette – Length-weight relationships. Northland College, Wisconsin, USA, 12 p.,http://www.ncfaculty.net/dogle/fishR/gnrlex/LengthWeight/LengthWeight.pdf.

Ogle D., 2011d. FishR vignette – Size structure analysis. Northland College, Wisconsin, USA, 54 p.,http://www.ncfaculty.net/dogle/fishR/gnrlex/VonBertalanffy/VonBertalanffy.pdf.

Pegg J. and Britton J.R., 2011. Effects of inter- and intra-specific competition on the growth rates ofjuvenile European barbel Barbus barbus used in the stock enhancement of UK fisheries. Fish.Res., 112, 8–12.

09p17

http://www.ncfaculty.net/dogle/fishR/gnrlex/SizeStructure/SizeStructure.pdf

http://www.ncfaculty.net/dogle/fishR/gnrlex/AgeComparisons/AgeComparisons.pdf

http://www.ncfaculty.net/dogle/fishR/gnrlex/LengthWeight/LengthWeight.pdf

http://www.ncfaculty.net/dogle/fishR/gnrlex/VonBertalanffy/VonBertalanffy.pdf


Penáz M. and Štouracová I., 1991. Effect of hydroelectric development on population dynamics ofBarbus barbus in the River Jihlava. Folia Zool., 40, 75–84.

Penáz M., Pivnicka K., Baruš V. and Prokeš M., 2003. Temporal changes in the abundance of barbel,Barbus barbus in the Jihlava River, Czech Republic. Folia Zool., 52, 441–448.

Penáz M., Svobodová Z., Baruš V., Prokeš M. and Drastichová J., 2005. Endocrine disruption in a barbel,Barbus barbus population from River Jihlava, Czech Republic. J. Appl. Ichthyol., 21, 420–428.

Persat H. and Chessel D., 1989. Typologie de distributions en classes de taille: intérêt dans l’étude despopulations de poissons et d’invertébrés. Acta Œcol., Œcol. Gen., 10, 175–195.

Pilcher M. and Copp G.H., 1997. Winter distribution and habitat use of fish in a regulated lowland riversystem of South-East England. Fish. Manag. Ecol., 4, 199–215.

Pilcher M., Copp G.H. and Szomolai V., 2004. A comparison of adjacent natural and channelisedstretches of a lowland river. Biologia-Bratislava, 59, 669–673.

Prokeš M., Šovcík P., Penáz M., Baruš V., Spurný P. and Vilizzi L., 2006. Growth of the barbel, Barbusbarbus, in the River Jihlava following major habitat alteration and estimated by two methods. FoliaZool., 55, 86–96.

Przybylski M., Boron A. and Kruk A., 2004. Growth of barbel, Barbus barbus (L.) in the upper WartaRiver, Odra River system. Ecohydrol. Hydrobiol. 2, 183–190.

R Development Core Team, 2010. R: A language and environment for statistical computing. RFoundation for Statistical Computing, Vienna, Austria. http://www.R-project.org.

Ricker W.E., 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish.Res. Board Can., 191.

Taylor A.A.L., Britton J.R. and Cowx I.G., 2004. Does the stock density of stillwater catch and releasefisheries affect the growth performance of introduced cultured barbel? J. Fish Biol., 65 (Suppl. A),308–313.

Tyler C.R. and Everett F., 1993. Incidences of gross morphological disorders in barbel (Barbus barbus)in three rivers in southern England. J. Fish Biol., 43, 739–748.

Vilizzi L. and Copp G.H., 2001. Behavioural responses of 0+ barbel in an artificial stream: distributionand velocity use. Animal Behav., 61, 645–654.

Vilizzi L. and Copp G.H., 2013a. Interstitial movement and emergence of barbel, Barbus barbus, freeembryos and larvae in a laboratory flume. J. Fish Biol., 82, 1057–1063.

Vilizzi L. and Copp G.H., 2013b. Bias, precision and validation of ageing 0+ European barbel Barbusbarbus (L.) from their otoliths. Cent. Eur. J. Biol. 8, 654–661.

Vilizzi L. and Walker K., 1999. Age and growth of the common carp, Cyprinus carpio, in the River Murray,Australia: validation, consistency of age interpretation, and growth models. Environ. Biol. Fish., 54,77–106.

Vilizzi L., Walker K.F., Jain T., McGlennon D. and Tsymbal V., 1998. Interpretability and precision ofannulus counts for calcified structures in carp, Cyprinus carpio L. Archiv Hydrobiol., 143, 121–127.

Vilizzi L., Copp G.H., Penáz M. and Carter M.G., 2006. Movement and abundance of barbel Barbusbarbus in a mesotrophic chalk stream: the River Lee, England. Folia Zool., 55, 183–197.

Watkins M.S., Doherty S. and Copp G.H., 1997. Microhabitat use by 0+ and older fishes in a smallEnglish chalk stream. J. Fish Biol., 50, 1010–1024.

Wheeler A. and Jordan D.R., 1990: The status of the barbel, Barbus barbus (L.) (Teleostei, cyprinidae),in the United Kingdom. J. Fish Biol., 37, 393–399.

Živkov M.T., Trichkova T.A. and Raikova-Petrova G.N., 1999. Biological reasons for the unsuitability ofgrowth parameters and indices for comparing fish growth. Environ. Biol. Fish., 54, 67–76.

09p18

http://www.R-project.org

age and growth of european barbel barbus barbus (cyprinidae) in

Documents